python EasyOCR库如何使用
说明
1、EasyOCR是一个用python编写的OCR三方库。可以在python中调用,用来识别图像中的文字,并输出为文本。
2、支持80多种语言的识别,识别精度高,甚至要超过PaddleOCR。
安装命令
pip install easyocr
代码实现
import easyocr #设置识别中英文两种语言reader = easyocr.Reader(["ch_sim","en"], gpu = False) # need to run only once to load model into memoryresult = reader.readtext(r"d:Desktop4A34A16F-6B12-4ffc-88C6-FC86E4DF6912.png", detail = 0)print(result)
实例扩展:
图文提取的代码
from pathlib import Pathimport easyocrfile_url = r"识别图片.jpg" # 需识别的图片split_symbol = " " # 默认空格为分隔符row_space = 15 # 默认字符高度为15px,当识别出来的字符间距超过这个数值时会换行。def make_reader(): # 将模型加载到内存中。模型文件地址 C:Users用户.EasyOCRmodel reader = easyocr.Reader(["ch_sim", "en"]) return readerdef change_to_character(file_url, reader, split_symbol=" ", row_space=15, save_dir="."): with open(file_url, "rb") as img: img_b = img.read() result = reader.readtext(img_b) result.sort(key=lambda x: x[0][0][1]) # 按竖直方向,进行排序==>进行分行处理。 # for i in result: # print(i) # print("="*100) # 按行进行分组 content = [] item = [result[0]] # 首先放入第一个元素 for i in result[1:]: if row_space >= i[0][0][1] - item[-1][0][0][1] >= 0: item.append(i) else: content.append(item) item = [i] content.append(item) filemane = Path(file_url).name.split(".")[0] with open(f"{save_dir}/{filemane}.txt", "w", encoding="utf8") as t: for i in content: # i 为每一行的内容 i.sort(key=lambda x: x[0][0][0]) # 对每行的内容进行先后排序 for r in i: # print(r) t.write(r[1] + split_symbol) t.write("") return contentif __name__ == "__main__": change_to_character(file_url, make_reader())
UI 界面的代码
import tkinter as tkfrom tkinter import filedialogfrom PIL import Image, ImageTkfrom pathlib import Pathfrom character import change_to_character, make_readerfrom threading import Threadimport time# class Showing(tk.Frame):# def __init__(self, master=None):# super().__init__(master)# self.master = master# self.pack()# # self.img = tk.PhotoImage(file=r"C:UsersyanhyDesktop捕获22.PNG")# self.create_widgets()## def create_widgets(self):# self.img = tk.PhotoImage(file=r"C:UsersyanhyDesktop捕获22.PNG")# self.img_wig = tk.Label(self, image=self.img)# self.img_wig.pack()# 最外层窗口设置root = tk.Tk()root.title("图片文字识别程序 联系:410889472@qq.com")window_x = root.winfo_screenwidth()window_y = root.winfo_screenheight()WIDTH = 1200HEIGHT = 750x = (window_x - WIDTH) / 2 # 水平居中y = (window_y - HEIGHT) / 3 # 垂直偏上root.geometry(f"{WIDTH}x{HEIGHT}+{int(x)}+{int(y)}")root.resizable(width=False, height=False)# 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》Row_space = 15File_url_list = []Img_type = [".jpg", ".jpeg", ".png", ".gif"]Split_symbol = " " # 间隔符。Save_dir = Path.cwd().joinpath("img_to_word")if Save_dir.is_dir(): passelse: Path.mkdir(Save_dir)# 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》def test(): print(f"{Row_space=}")def choose_file(): # 获取导入的图片路径地址 global show_img, img_label, text, File_url_list filenames = filedialog.askopenfilenames() if len(filenames) == 1 and len(File_url_list) == 0: # 单张图片导入,显示图片 if Path(filenames[0]).suffix.lower() in Img_type: # 判断是否图片类型 File_url_list = list(filenames) try: if text.winfo_exists(): text.destroy() except NameError as e: print(f"choose_file提示:张图片导入错误>>> {e}") try: if img_label.winfo_exists(): img_label.destroy() except NameError as e: print(f"choose_file提示:单张图片导入错误>>> {e}") img = Image.open(File_url_list[0]).resize((560, 660)) # print(img.size) show_img = ImageTk.PhotoImage(image=img) img_label = tk.Label(f_left, image=show_img) img_label.pack() else: print("导入的是非图像格式") else: # 多张图片导入,显示列表。 try: if img_label.winfo_exists(): img_label.destroy() except NameError as e: print(f"提示:多张图片导入错误>>> {e}") try: if text.winfo_exists(): text.destroy() except NameError as e: print(f"提示:多张图片导入错误>>> {e}") text = tk.Text(f_left, spacing1=5, spacing3=5) text.pack(fill="both", expand=True) for i in filenames: if Path(i).suffix.lower() in Img_type: File_url_list.append(i) else: pass File_url_list = set(File_url_list) for i in list(File_url_list): # 把文件写入到文本框中 text.insert("end", str(list(File_url_list).index(i)+1) + ": " + i + "") File_url_list = list(File_url_list) print(f"{File_url_list=}")def choose_dir(): global show_img, img_label, text, File_url_list directoryname = filedialog.askdirectory() print(f"{directoryname=}") try: if img_label.winfo_exists(): img_label.destroy() except NameError as e: print(f"choose_dir提示:多张图片导入错误>>> {e}") try: if text.winfo_exists(): text.destroy() except NameError as e: print(f"choose_dir提示:多张图片导入错误>>> {e}") text = tk.Text(f_left, spacing1=5, spacing3=5) text.pack(fill="both", expand=True) for i in Path(directoryname).iterdir(): # 获取文件夹下的所有文件。 if Path(i).suffix.lower() in Img_type: File_url_list.append(i.as_posix()) # as_posix() 把Path型转为字符串。 else: pass File_url_list = set(File_url_list) for i in list(File_url_list): # 把文件写入到文本框中 text.insert("end", str(list(File_url_list).index(i) + 1) + ": " + i + "") File_url_list = list(File_url_list) print(f"{File_url_list=}")def clear_file_list(): global File_url_list File_url_list.clear() try: if img_label.winfo_exists(): img_label.destroy() except NameError as e: print(f"clear_file_list提示:清空错误>>> {e}") try: if text.winfo_exists(): text.destroy() except NameError as e: print(f"clear_file_list提示:清空错误错误>>> {e}")def get_entry1(): # 设置换行间距变量值 global Row_space num = entry1.get() if num.isdigit(): if int(num) > 0: Row_space = int(num) else: entry1.delete(0, "end") entry1.insert(0, 15) Row_space = 15def set_split_symbol(): global Split_symbol Split_symbol = entry2.get() print(f"{Split_symbol=}")def do_change(): if File_url_list: v.set("文字提取中,请稍后……") button_do.config(state="disable") # 使按钮不可用。 # ======================================== def main(): reader = make_reader() for i in File_url_list: content = change_to_character(i, reader, row_space=Row_space, split_symbol=Split_symbol, save_dir=Save_dir) read_text.delete(1.0, "end") for c in content: # i 为每一行的内容 c.sort(key=lambda x: x[0][0][0]) # 对每行的内容进行先后排序 for r in c: # print(r) read_text.insert("end", r[1] + Split_symbol) read_text.insert("end", "") v.set("文字提取结束。") button_do.config(state="normal") # 恢复按钮可用。 # ======================================== t = Thread(target=main, daemon=True) t.start() else: v.set("请先选择图片!")def join_file(): v.set("文件开始合并。") filst = list(Path(Save_dir).iterdir()) # 获取文件夹中所有的文本文件。 with open(f"{Save_dir}/合并文件.txt", "w", encoding="utf8") as join_f: for f in filst: with open(f, "r", encoding="utf8") as r_f: read_con = r_f.read() join_f.write(f.name+""+read_con + "") time.sleep(1) v.set("文件合并完毕。")# 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》f_top = tk.Frame(root, height=65, width=1100, bd=1, relief="flat") # "sunken" "raised","groove" 或 "ridge"f_top.pack_propagate(False) # 如果不加这个参数,当Frame框架中加入部件时,会自动变成底层窗口,自身的特性会消失。f_top.pack(side="top", pady=5)f_left = tk.Frame(root, height=660, width=560, bd=1, relief="groove")f_left.pack_propagate(False)f_left.pack(side="left", padx=20)f_right = tk.Frame(root, height=660, width=560, bd=1, relief="groove")f_right.pack_propagate(False)f_right.pack(side="left", padx=20)read_text = tk.Text(f_right, spacing1=5, spacing3=5)read_text.pack(fill="both", expand=True)# 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》button_choose_file = tk.Button(f_top, text="选择图片", command=choose_file)button_choose_file.pack(side="left", padx=10, ipadx=5)button_choose_file = tk.Button(f_top, text="选择文件夹", command=choose_dir)button_choose_file.pack(side="left", padx=10, ipadx=5)button_clear_file = tk.Button(f_top, text="清空选择", bg="#FFEF2F", command=clear_file_list)button_clear_file.pack(side="left", padx=5, ipadx=5)# 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》f_row_content = tk.Frame(f_top, height=50, width=300, bg="#D1D4D0", relief="flat") # "sunken" "raised","groove" 或 "ridge"f_row_content.pack_propagate(False)f_row_content.pack(side="left", padx=15)button_set_row_height = tk.Button(f_row_content, text="设置行间距", command=get_entry1)button_set_row_height.pack(side="left", ipadx=3, padx=3)entry1 = tk.Entry(f_row_content, font=("", 18), width=3)entry1.insert(0, 15)entry1.pack(padx=5, side="left")tk.Label(f_row_content, justify="left", text="填入像素值,设置换行间距。默认15个像素。").pack(side="left")# 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》f_split = tk.Frame(f_top, height=50, width=215, bg="#D1D4D0", relief="flat") # "sunken" "raised","groove" 或 "ridge"f_split.pack_propagate(False)f_split.pack(side="left", padx=4)button_split = tk.Button(f_split, text="设置分隔符", command=set_split_symbol)button_split.pack(side="left", ipadx=3, padx=3)entry2 = tk.Entry(f_split, font=("", 18), width=3)entry2.insert(0, " ")entry2.pack(padx=5, side="left")tk.Label(f_split, justify="left", text="默认一个空格").pack(side="left")# 《《《《《《《《《《《《《《《《《《《《《《 提取 合并文件 》》》》》》》》》》》》》》》》》》》》》》》》》button_do = tk.Button(f_top, text="开始提取", bg="#4AB0FF", command=do_change)button_do.pack(side="left", padx=10, ipadx=2)button_join = tk.Button(f_top, text="合并文件", command=join_file)button_join.pack(side="left", padx=5, ipadx=2)v = tk.StringVar()v.set("info……")tk.Label(f_top, bg="#2EBD1D", justify="left", textvariable=v).pack(side="left")# 《《《《《《《《《《《《《《《《《《《《《《 右键菜单 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》def copy_text(): read_text.event_generate(">")menubar = tk.Menu(tearoff=False)# root["menu"] = menubar # 没有把这个 菜单部件 加入到 root 窗口的菜单属性中,所以它不会在root窗口的顶部显示。menubar.add_command(label="复制", command=copy_text)def show_menu(event): """用 菜单部件 的 post 方法展示菜单""" menubar.post(event.x_root, event.y_root)read_text.bind("<button-3>", show_menu)# 》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》》root.mainloop()</button-3>